The present invention generally relates to storage control in a file storage system having a plurality of file storage apparatuses.
Generally, a file storage system has two file storage apparatuses (“file storages,” hereinafter) related one-on-one to each other (Japanese Patent Application Publication No. 2006-164211). A first file storage migrates a file to be migrated, to a second file storage related to the first file storage, and generates a stub for the file to be migrated. The term “stub” means an object (e.g., metadatum) associated with information on a destination where the file is stored (information on a link). Upon reception of a read command on the file related to the stub from a client, the first file storage transmits the read command according to the link of the stub to the second file storage that has a file related to the stub. Consequently, the first file storage acquires the file related to the stub from the second file storage, and transmits this file to the client.
When migrated, the file may pass through a migration path. There are two nodes configuring the migration path. Each of the nodes may be a unit of storage. For example, one of the nodes may serve as a file system managed by the first file storage, and the other node may serve as a file system managed by the second file storage.
There is known a cloud storage, a storage system that provides an unspecified number of users with functions of a storage via a network. The cloud storage is often configured by a plurality of file storages.
When adopting the cloud storage as at least either a file migration source or a file migration destination, a file storage system might have three or more file storages.
In such file storage system, another node might be required as an intermediate node of the migration path, in addition to the both end nodes (the two nodes). According to such migration path, three or more file systems of the three or more file storages are in a cascade form, wherein files are migrated from a front end node (file system) to a terminal end node (file system) via the intermediate node (file system). The conventional file storage system described above, however, is configured only by the two file storages that are related one-on-one to each other. It is therefore difficult, according to the prior art, to construct a migration path having such intermediate node.
It is considered that the higher the number of file storages configuring a file storage system, the more complicated the configuration of a migration path and the higher the number of migration paths. Therefore, the management burden on an administrator of the file storage system is considered enormous. It is difficult to investigate, for example, what kind of file storage the files have passed through previously.
Such problems could happen not only in the file storage system having the cloud storage, but also in other file storage systems having three or more file storages.